-
Notifications
You must be signed in to change notification settings - Fork 2
/
Function.py
265 lines (258 loc) · 11.5 KB
/
Function.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
import numpy as np
# import matplotlib.pyplot as plt
from operator import truediv
import csv,threading
from sklearn import manifold
import torch.nn.functional as F
import torch
import scipy
# import cv2
def get_k_layer_feature_map(model_layer, k, x):
with torch.no_grad():
for index, layer in enumerate(model_layer):#model的第一个Sequential()是有多层，所以遍历
if index==5:
continue
x = layer(x)#torch.Size([1, 64, 55, 55])生成了64个通道
if index==4:
x = torch.squeeze(x, 2)
if k == index:
return x
def show_feature_map(feature_map, feature_data, i):
feature_map = feature_map.squeeze(0) # 压缩成torch.Size([64, 55, 55])
# 现在4维，包括一个batch，测试的时候可以把batch去掉
# 以下4行，通过双线性插值的方式改变保存图像的大小
unsample = torch.nn.UpsamplingBilinear2d(size=(64, 64))
feature_data = torch.sum(feature_data, 2)
# feature_data = feature_data.view(1, feature_data.shape[2],
# feature_data.shape[3], feature_data.shape[4]) # (1,64,55,55)
# feature_data = torch.sum(feature_data, 1)
feature_data = unsample(feature_data)
feature_data = np.array(feature_data.cpu())
# upsample = torch.nn.UpsamplingBilinear2d(size=(256, 256)) # 这里进行调整大小
# feature_map = upsample(feature_map)
# feature_map = feature_map.view(feature_map.shape[1], feature_map.shape[2],
# feature_map.shape[3])
feature_data = feature_data.squeeze(0)
feature_data = feature_data.squeeze(0)
feature_map = torch.sum(feature_map, 0)
feature_map = feature_map.view(1, 1,feature_map.shape[0], feature_map.shape[1]) # (1,64,55,55)
feature_map = unsample(feature_map)
feature_map = np.array(feature_map.cpu())
feature_map = feature_map.squeeze(0)
feature_map = feature_map.squeeze(0)
img_add = cv2.addWeighted(feature_data, 0.7, feature_map, 0.3, 0)
# feature_map_num = feature_map.shape[0] # 返回通道数
# row_num = np.ceil(np.sqrt(feature_map_num)) # 8
plt.figure()
plt.imshow(img_add, cmap='jet') # feature_map[0].shape=torch.Size([55, 55])
plt.axis('off')
plt.show()
plt.imsave('./feature_map_save/'+ str(i)+'.svg', img_add)
plt.imsave('./feature_map_save/'+ str(i)+'_origin.svg', feature_data)
# for index in range(1, feature_map_num + 1): # 通过遍历的方式，将64个通道的tensor拿出
# plt.subplot(row_num, row_num, index)
# plt.imshow(feature_map[index - 1].cpu(), cmap='jet') # feature_map[0].shape=torch.Size([55, 55])
# # 将上行代码替换成，可显示彩色 plt.imshow(transforms.ToPILImage()(feature_map[index - 1]))#feature_map[0].shape=torch.Size([55, 55])
# plt.axis('off')
# # scipy.imsave('./feature_map_save/' + str(index) + ".png", feature_map[index - 1].cpu())
# plt.show()
def DrawCluster(label, cluster, oa):
label = np.array(label)
palette = np.array([[0, 139, 0],
[0, 0, 255],
[255, 255, 0],
[255, 127, 80],
[255, 0, 255],
[139, 139, 0],
[0, 139, 139],
[0, 255, 0],
[0, 255, 255],
[0, 30, 190],
[127, 255, 0],
[218, 112, 214],
[46, 139, 87],
[0, 0, 139],
[255, 165, 0],
[127, 255, 212],
[218, 112, 214],
[255, 0, 0],
[205, 0, 0],
[139, 0, 0],
[65, 105, 225],
[240, 230, 140],
[244, 164, 96]])
# palette = np.array([[0, 0, 255],
# [76, 230, 0],
# [255, 190, 232],
# [255, 0, 0],
# [156, 156, 156],
# [255, 255, 115],
# [0, 255, 197],
# [132, 0, 168],
# [0, 0, 0]])
palette = palette * 1.0 / 255
tsne = manifold.TSNE(n_components=2,init='pca')
X_tsne = tsne.fit_transform(cluster)
x_min, x_max = X_tsne.min(0), X_tsne.max(0)
X_norm = (X_tsne-x_min)/(x_max-x_min)
plt.figure()
for i in range(23):
xx1 = X_norm[np.where(label==i), 0]
yy1 = X_norm[np.where(label==i), 1]
plt.scatter(xx1,yy1, color=palette[i].reshape(1,-1),s=20, linewidths=2)
plt.xlim(np.min(X_norm)-0.0001, np.max(X_norm)+0.0001)
plt.ylim(np.min(X_norm)-0.0001, np.max(X_norm)+0.0001)
# plt.legend(['Brocoli_green_weeds_1', 'Brocoli_green_weeds_2', 'Fallow', 'Fallow_rough_plow',
# 'Fallow_smooth',
# 'Stubble', 'Celery', 'Grapes_untrained', 'Soil_vinyard_develop', 'Corn_senesced_green_weeds',
# 'Lettuce_romaine_4wk', 'Lettuce_romaine_5wk', 'Lettuce_romaine_6wk', 'Lettuce_romaine_7wk',
# 'Vinyard_untrained', 'Vinyard_vertical_trellis'], bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0)
plt.savefig('./CLUSTER/' + 'HHK'+ str('.%6f'%oa)+'.svg',dpi=600, bbox_inches='tight')
plt.show()
# plt.savefig("./cluster.png")
# Y = tsne
def DrawResult(y_pred, imageID, OA):
# ID=1:Pavia University
# ID=2:Indian Pines
# ID=6:KSC
# labels = labels + 1
num_class = y_pred.max()
if imageID == 'PU':
row = 610
col = 340
palette = np.array([[0, 0, 255],
[76, 230, 0],
[255, 190, 232],
[255, 0, 0],
[156, 156, 156],
[255, 255, 115],
[0, 255, 197],
[132, 0, 168],
[0, 0, 0]])
palette = palette * 1.0 / 255
elif imageID == 'IP':
print('?')
row = 145
col = 145
palette = np.array([[0, 168, 132],
[76, 0, 115],
[0, 0, 0],
[190, 255, 232],
[255, 0, 0],
[115, 0, 0],
[205, 205, 102],
[137, 90, 68],
[215, 158, 158],
[255, 115, 223],
[0, 0, 255],
[156, 156, 156],
[115, 223, 255],
[0, 255, 0],
[255, 255, 0],
[255, 170, 0]])
palette = palette * 1.0 / 255
elif imageID == 'SA':
row = 512
col = 217
palette = np.array([[0, 168, 132],
[76, 0, 115],
[0, 0, 0],
[190, 255, 232],
[255, 0, 0],
[115, 0, 0],
[205, 205, 102],
[137, 90, 68],
[215, 158, 158],
[255, 115, 223],
[0, 0, 255],
[156, 156, 156],
[115, 223, 255],
[0, 255, 0],
[255, 255, 0],
[255, 170, 0]])
palette = palette * 1.0 / 255
elif imageID == 'PD':
row = 377
col = 512
palette = np.array([[237, 227, 81],
[167, 237, 81],
[0, 0, 0],
[181, 117, 14],
[77, 122, 15],
[186, 186, 186]])
palette = palette * 1.0 / 255
elif imageID == 'HHK':
row = 1147
col = 1600
palette = np.array([[0, 139, 0],
[0, 0, 255],
[255, 255, 0],
[255, 127, 80],
[255, 0, 255],
[139, 139, 0],
[0, 139, 139],
[0, 255, 0],
[0, 255, 255],
[0, 30, 190],
[127, 255, 0],
[218, 112, 214],
[46, 139, 87],
[0, 0, 139],
[255, 165, 0],
[127, 255, 212],
[218, 112, 214],
[255, 0, 0],
[205, 0, 0],
[139, 0, 0],
[65, 105, 225],
[240, 230, 140],
[244, 164, 96]])
palette = palette * 1.0 / 255
X_result = np.zeros((y_pred.shape[0],3))
for i in range(0, num_class+1):
X_result[np.where(y_pred == i), 0] = palette[i, 0]
X_result[np.where(y_pred == i), 1] = palette[i, 1]
X_result[np.where(y_pred == i), 2] = palette[i, 2]
X_result = np.reshape(X_result, (row, col, 3))
# X_mask[1:-1,1:-1,:] = X_result
plt.axis("off")
plt.imsave('./image_result/'+imageID + '_s3net' + str(OA) + '.svg',X_result)
return X_result
def neibor_result_choose(y_pred):
y_pred_count = []
new_y_pred = []
for i in range(len(y_pred)):
y_pred_count.append(y_pred[i])
if (i+1) % 25 == 0:
counts = np.bincount(y_pred_count)
index = np.argmax(counts)
max_time = np.max(counts)
counts[index] = 0
if max_time in counts:
new_y_pred.append(np.argmax(y_pred_count[4]))
else:
new_y_pred.append(index)
y_pred_count.clear()
y_pred_count = []
return new_y_pred
def AA_andEachClassAccuracy(confusion_matrix):
counter = confusion_matrix.shape[0]
list_diag = np.diag(confusion_matrix)
list_raw_sum = np.sum(confusion_matrix, axis=1)
each_acc = np.nan_to_num(truediv(list_diag, list_raw_sum))
average_acc = np.mean(each_acc)
return each_acc, average_acc
def information_process(dataset, windowSize1, windowSize2, perclass, batch_size, iteration, K, add_info, Each_class_acc, margin):
res = []
for i in range(len(np.mean(Each_class_acc, 0))):
str_ = str('%.2f' % np.mean(Each_class_acc, 0)[i]) + '+-' + str('%.2f' % np.std(Each_class_acc, 0)[i])
res.append(str_)
infomation = [dataset, 'windowSize1:',windowSize1,'windowSize2',windowSize2, 'perclass:', perclass, 'margin:', margin, 'iteration:', iteration, 'PCA:',K, 'batch_size:',batch_size, 'oa:',np.mean(add_info, 0)[0],
'+-', np.std(add_info, 0)[0], 'kappa:', np.mean(add_info, 0)[1], '+-', np.std(add_info, 0)[1], 'aa:',np.mean(add_info, 0)[2], '+-', np.std(add_info, 0)[2]
,'each_acc:', res]
print('oa:', np.mean(add_info, 0)[0], '+-', np.std(add_info, 0)[0], 'kappa:', np.mean(add_info, 0)[1], '+-',
np.std(add_info, 0)[1], 'aa:', np.mean(add_info, 0)[2], '+-', np.std(add_info, 0)[2])
csvFile = open("./Final_Experiment.csv", "a")
writer = csv.writer(csvFile)
writer.writerow(infomation)
csvFile.close()